Using Dynamic Simulations to Improve Geological Models: A Full Field Integrated Study of an HPHT Gas Field in the Norwegian Sea

Abstract

Abstract Assisted history matching is a pivotal process in optimizing reservoirs models through the calibration to observed well and pressure field data. Robust reservoir forecasting is essential for effective asset management and development planning; however, it is significantly impacted by uncertainties inherent in the subsurface. Large uncertainties exist in the underlying geological and simulation inputs which pose a key challenge to achieve an accurate and meaningful match between observed and simulated data; especially when considering complex, marginal or high environment impact projects in today's volatile and unpredictable financial landscape. Therefore, there is a prevailing need for fit-for-purpose models that encompass, assess and mitigate pivotal uncertainties. These models play a crucial role in informing management decisions, reducing the margin for error, and mitigating potential financial setbacks. The current study is based on the geological and dynamic understanding of the Fogelberg discovery, a complex tidal bar system in the Norwegian Sea. Available BHP measurements obtained during a DST test were the main source of data to calibrate the model, refine its geological understanding and update the static and dynamic model. The main objective of creating a fit-for-purpose model was to make informed decisions on further development stages to maximize ultimate recovery factor of an asset understood to have connectivity complexities and uncertainties. A range of different development scenarios and their complexity were modelled, including developing the field with hydraulically fractured wells, complete wells with Fishbones or slotted liner. Complexities and uncertainties in terms of reservoir geometry, properties distribution and overall connectivity were critical to determine the commercial feasibility of the asset development. Hence, a full multi-scale uncertainty ensemble model covering the placement of low permeability inter-bars to stylolitisation at core-plug scale was key to achieve a robust DST match while honouring the spatial geometry and observed tidal bar complexity. The static model was completed using a scenario-based approach (after Bentley and Smith, 2008). The depositional concept model was honoured while allowing modelling flexibility for connectivity uncertainties. The DST data from well 6506-9/4A provided indication of the reservoir connectivity seen around the well, however, it was crucial not to make changes based on these potentially local results that would be spread at the field level. Based on initial results from the dynamic simulation, further modifications on the geological variograms were made, marrying the analysis from the DST and the geological understanding of the field in an iterative feedback loop between static and dynamic domains. The integrated static-dynamic model produced a robust match to the bottom hole pressure (BHP) reported during the DST and was used to evaluate the asset production potential. A direct comparison between the recoverable volumes obtained from wells placed in the same position but completed with different technologies was made to permit evaluation of development strategies including the use fishbones and hydraulic fractured completions. The focus of this paper lies in emphasizing the significance of uncertainty management by employing geological and dynamic scenarios to maximize recovery factor and evaluate all possible aspects to make the development financially attractive while adopting a multi-scale approach to precisely depict the varying scales at which heterogeneities exist. This methodology contributes to the reliability and predictive capability of surface reservoir models, through optimizing reservoir performance and recovery strategies under uncertainty and enhancing our understanding of subsurface and the interplay of static and dynamic models.